Analyzing scale independence in jobs-housing and commute efficiency metrics

Understanding journey to work travel patterns remains an important concern for planners and policy-makers from the viewpoint of economic, environmental, and social sustainability. Researchers, keen to inform metropolitan scale planning efforts, have devised ways of benchmarking regional commuting and land use phenomena. The foundation for these benchmarks rests on metrics that quantify the home-job proximity in terms of the aggregate arrangement of workers relative to jobs. Emanating from the literature on 'excess commuting' and 'jobs housing balance', these metrics are increasingly moving towards policy applications. Despite major methodological developments over the last decade, a key methodological issue remains unresolved. Recently developed metrics under this regional macro-scale framework use zonal-based spatial data (e.g. census tracts or traffic analysis zones (TAZs)) and consequently the values of the metrics may be influenced by the scale (e.g. zone size varies between census blocks versus tracts) and zonal partitioning scheme. Moreover it is not known if values of these metrics vary across scale, and exhibit self-similarity, meaning whether it is possible to infer values from one scale to another. This study examines the relationship between the commuting efficiency framework and spatial scale issues by implementing a suite of commuting metrics in the Boise, Idaho USA metropolitan area. Simulations using geographic information systems (GIS), optimization techniques and fractal analysis show that newer metrics developed post 2002 do not vary with scale, while those devised pre-2002 vary with scale but do so in a predictable way.